Smart Rivers: AI’s Role in Tracking Pollution

Artificial intelligence's analytical capabilities make it a powerful tool for tackling complex issues that involve vast amounts of data, such as climate analysis. Now, AI is emerging as a key player in environmental protection, as demonstrated by one of its latest applications. Unregulated discharges into rivers remain a persistent challenge in water management. Detecting them promptly and tracing their origin is crucial for swift decision-making. In the UK—where such incidents have become increasingly frequent—AI, supported by volunteer efforts, promises to accelerate and improve response measures.

The challenge of river pollution in the UK

According to a recent report in The Guardian, untreated sewage discharges in England and Wales have reached alarming levels, totalling 3.6 million hours in 2023—more than double the figure recorded the previous year. The newspaper attributes this to inadequate investment in infrastructure and lenient regulation, resulting in large-scale contamination of rivers and coastal waters.

Against this backdrop, the Water Research Centre (WRC), a UK-based scientific organisation, has developed an artificial intelligence system aimed at protecting the country’s environment and natural resources.

AI in the service of environmental protection

Traditionally, monitoring river water quality has relied on costly equipment and chemical analysis conducted by regulatory bodies and private companies. This project, however, offers a more accessible and cost-effective alternative—leveraging AI-driven image analysis.

Developed in collaboration with the WRC, National Taiwan University, and a rainfall modelling company, the initiative is built on a vast collection of photographs taken by volunteers. Specifically, more than 1,000 members of the Friends of Bradford’s Becks community contributed images documenting visual indicators of river health in the Bradford Beck catchment. These images were used to train an AI system capable of identifying signs of pollution.

The local community’s involvement in the project stems from a severe pollution incident in 2018 that devastated aquatic life in the river, with lasting effects over two years. Since then, the group has explored remote monitoring solutions to detect and report new pollution events.

During the project's initial six-month phase in 2024, researchers defined key visual indicators of river health, such as the presence of wildlife and vegetation, changes in water colour, litter accumulation, obstacles in the riverbed, and runoff discharges.

The AI models trained on these images have delivered promising results. Systems like C-Tran and ChatGPT demonstrated high accuracy in identifying key pollution indicators, while YOLOv8, though less precise overall, proved useful for visually pinpointing contamination hotspots. The AI also detected drainage points near sanitary waste, highlighting priority areas for further investigation.

Natural England, which funded the project, has confirmed plans to refine the AI models to enhance their accuracy and expand their application in water quality monitoring.

Other AI initiatives for river health

The work at Bradford Beck is not the only AI-driven initiative aimed at improving river health in the UK. Recently, the consultancy firm Capgemini partnered with an organisation focused on safeguarding Britain’s waterways using AI and Big Data technologies. This project integrates multiple data sources, including sensor-based monitoring, to assess the most affected areas and predict potential issues before they escalate.

If you’re interested in AI’s environmental applications, you might also want to explore how innovative technologies are transforming water treatment and desalination—such as this system for generating biogas from wastewater.

Source:

• The Guardian
• Aquatechtrade